Construcción y análisis de los coeficientes de sendero
Abstract
El presente artículo contiene una explicación de los Coeficientes de Sendero, desde un punto de vista matemático-estadístico. Este método es bastante útil para determinar relaciones efecto-causa, consiste en realizar un análisis estadístico de causa y efecto en variables que se encuentran correlacionadas, su análisis tiene como meta expresar una variable dependiente ' y' en función de efectos directos e indirectos de variables independientes xi. Cuando se estudian las relaciones que existen entre una variable dependiente 'y' y un conjunto de variables independientes xi, generalmente se utiliza un análisis de regresión y/o correlación, no obstante, el análisis de coeficientes de sendero, es un método que permite analizar la interdependencia entre dichas variables; empleando de este modo, la regresión y correlación, sólo de forma complementaria. Por ejemplo, la selección indirecta de variables relacionadas con una variable de respuesta, necesita la identificación de características simples y altamente asociadas con la variable dependiente. Esta identificación se basa generalmente en el análisis de correlación; que determina un índice (coeficiente de correlación) o referencia acerca de la relación entre las variables, pero este análisis es restringido en el sentido de que sólo brinda información entre variables una a una, vale decir, que es una información entre pares de variables, entonces muchas características que aparentemente no tienen relación con la variable dependiente, se debe a que los efectos de las variables independientes no son directos; sino que se relacionan indirectamente y el análisis de coeficientes de sendero, es una técnica bastante útil para determinar dichas relaciones de efecto-causa y la magnitud de dichos coeficientes; precisamente brindan información de la relación, en función de efectos directos e indirectos. Este método es muy conocido en el campo de la agronomía y ya fue empleado en diversos cultivos, por tanto, a manera de ejemplo se verá su aplicación brevemente sobre el cultivo del tubérculo ajipa (pachiryzus ajipa). Entonces, se realiza primero un análisis de regresión sobre el rendimiento de vainas en cultivos de ajipa, rescatando las variables estadísticamente más significativas (a un nivel de 0,05), y posteriormente se aplica el análisis de correlación con las variables restantes ya que ayudará en la interpretación final corroborando las relaciones existentes y de este modo ratificar la interrelación mediante el análisis de sendero. Los resultados que ofrece este análisis, serán más precisos dentro de la aplicación del problema.
The present article contains an explanation of the Path coefficients, from a mathematical-statistical point of view. This method is quite useful to determine effect-cause relationships, it consists of performing a statistical analysis of cause and effect in variables that are correlated, its analysis aims to express a dependent variable " y" as a function of direct and indirect effects of variables independent "x¡". When studying the relationships that exist between a dependent variable 'y' and a set of independent variables 'x¡', a regression and / or correlation analysis is generally used, however, the analysis of path coefficients is a method that allows to analyze the interdependence between these variables; using in this way, the regression and correlation, only in a complementary way. For example, the indirect selection of variables related to a response variable requires the identification of simple characteristics highly associated with the dependent variable. This identification is usually based on the correlation analysis; which determines an index (correlation coefficient) or reference about the relationship between variables, but this analysis is restricted in the sense that it only provides information between variables one by one it means that, it is information between pairs of variables, so, many characteristics that apparently have no relation with the dependent variable, is due to the fact that the effects of the independent variables are not direct; but they are related indirectly and the analysis of path coefficients, is a very useful technique to determine these effect-cause relationships and the magnitude of said coefficients; they precisely provide information on the relationship, based on direct and indirect effects. This method is well known in the field of agronomy and it was already used in several crops, so, as an example its application will be briefly on the crops of the ajipa tuber (pachiryzus ajipa). Then, a regression analysis is performed first on the yield of pods in ajipa crops, rescuing the most statistically significant variables (at level of 0.05), and subsequently the correlation analysis with the remaining variables is applied since it will help in the final interpretation, corroborating the existing relationships and in this way validating the interrelation through the path analysis. The results offered by this analysis will be more precise at the application of the problem.
The present article contains an explanation of the Path coefficients, from a mathematical-statistical point of view. This method is quite useful to determine effect-cause relationships, it consists of performing a statistical analysis of cause and effect in variables that are correlated, its analysis aims to express a dependent variable " y" as a function of direct and indirect effects of variables independent "x¡". When studying the relationships that exist between a dependent variable 'y' and a set of independent variables 'x¡', a regression and / or correlation analysis is generally used, however, the analysis of path coefficients is a method that allows to analyze the interdependence between these variables; using in this way, the regression and correlation, only in a complementary way. For example, the indirect selection of variables related to a response variable requires the identification of simple characteristics highly associated with the dependent variable. This identification is usually based on the correlation analysis; which determines an index (correlation coefficient) or reference about the relationship between variables, but this analysis is restricted in the sense that it only provides information between variables one by one it means that, it is information between pairs of variables, so, many characteristics that apparently have no relation with the dependent variable, is due to the fact that the effects of the independent variables are not direct; but they are related indirectly and the analysis of path coefficients, is a very useful technique to determine these effect-cause relationships and the magnitude of said coefficients; they precisely provide information on the relationship, based on direct and indirect effects. This method is well known in the field of agronomy and it was already used in several crops, so, as an example its application will be briefly on the crops of the ajipa tuber (pachiryzus ajipa). Then, a regression analysis is performed first on the yield of pods in ajipa crops, rescuing the most statistically significant variables (at level of 0.05), and subsequently the correlation analysis with the remaining variables is applied since it will help in the final interpretation, corroborating the existing relationships and in this way validating the interrelation through the path analysis. The results offered by this analysis will be more precise at the application of the problem.
Description
Vol. 8, No. 4